Home › Categories › Business & Finance› ссылка

ссылка.pdf

2102.12092.pdf

🔗 Source: arxiv.org

📊 Size: 9.66 MB

👤 Author: Aditya Ramesh, Mikhail Pavlov, Gabriel Goh, Scott Gray, Chelsea Voss, Alec Radford, Mark Chen, Ilya

⬇️ Downloads: 182

Summary

Text-to-image generation has traditionally focused on finding better modeling assumptions for training on a fixed dataset, involving complex architectures, auxiliary losses, or side information. A simple approach is proposed, based on a transformer that autoregressively models text and image tokens as a single stream of data. With sufficient data and scale, this approach is competitive with previous domain-specific models when evaluated in a zero-shot fashion. Training a 12-billion parameter autoregressive transformer on 250 million image-text pairs results in a flexible, high-fidelity generative model of images controllable through natural language. The system achieves high-quality image generation on the MS-COCO dataset zero-shot, without using training labels, and is preferred over prior work by human evaluators 90% of the time.

Download File

Description

Text-to-image generation has traditionally focused on finding better modeling assumptions for training on a fixed dataset, involving complex architectures,...

Technical Information

File Format: PDF
File Size: 9.66 MB
Pages: 20
Language: EN
Author: Aditya Ramesh, Mikhail Pavlov, Gabriel Goh, Scott Gray, Chelsea Voss, Alec Radford, Mark Chen, Ilya
Total Downloads: 182
Last Updated: 1 month ago

Document Overview

This PDF document about ссылка provides comprehensive information and guidance. Whether you're a beginner or advanced user, this resource offers valuable insights into ссылка.

ссылка.pdf

Summary

Description

Technical Information

Document Overview

Related Topics

Related Categories

Related Documents

ссылка

ссылка

ссылка

ссылка

ссылка